Microsoft’s new AI ‘code of conduct’ tells models not to hack systems or trick humans
Microsoft’s AI code of conduct defines broad principles and concrete safety constraints to ensure its models support people, promote human flourishing, and avoid replacing human roles.
MAIN POINTS
- Microsoft AI models should support humans rather than replace them.
- The code emphasizes accelerating human flourishing as a core principle.
- General principles are paired with specific safety constraints.
- Safety rules are intended to operationalize the stated principles.
TAKEAWAYS
- The policy frames AI as a tool for human assistance, not substitution.
- Human flourishing is treated as a central design goal.
- Principles alone are not enough; concrete constraints enforce them.
- The code connects ethical intent with practical safety implementation.